An empirically derived approach to the classification and diagnosis of mood disorders

Abstract

This article describes a system for diagnosing mood disorders that is empirically derived and designed for its clinical utility in everyday practice. A random national sample of psychiatrists and clinical psychologists described a randomly selected current patient with a measure designed for clinically experienced informants, the Mood Disorder Diagnostic Questionnaire (MDDQ), and completed additional research forms. We applied factor analysis to the MDDQ to identify naturally occurring diagnostic groupings within the patient sample. The analysis yielded three clinically distinct mood disorder dimensions or spectra, consistent with the major mood disturbances included in the DSM and ICD over successive editions (major depression, dysthymia, and mania), along with a suicide risk index. Diagnostic criteria were determined strictly empirically. Initial data using diagnostic efficiency statistics supported the accuracy of the dimensions in discriminating DSM-IV diagnoses; regression analyses supported the discriminant validity of the MDDQ scales; and correlational analysis demonstrated coherent patterns of association with family history of mood disorders and functional outcomes, supporting validity. Perhaps most importantly, the MDDQ diagnostic scales demonstrated incremental validity in predicting adaptive functioning and psychiatric history over and above DSM-IV diagnosis. The empirically derived syndromes can be used to diagnose mood syndromes dimensionally without complex diagnostic algorithms or can be combined into diagnostic prototypes that eliminate the need for everexpanding categories of mood disorders that are clinically unwieldy.